Magnetically-controlled arc fuse additive manufacturing system and manufacturing method thereof

Abstract

The invention belongs to the technical field of arc fuse additive manufacturing, and particularly discloses a magnetic control arc fuse additive manufacturing system and a manufacturing method thereof.The manufacturing system comprises a CMT welding machine, a magnetic generation device, a molten pool image detection device and a computer, and the magnetic generation device comprises an electromagnetic generation part and a power supply part; the electromagnetic generation part comprises an annular iron core, a coil, a first U-shaped iron core and a second U-shaped iron core, the power supply part is electrically connected with the coil, and after the coil is powered on, the annular iron core can generate an axial magnetic field in the axis direction of a welding wire at the welding wire; the U-shaped opening of the first U-shaped iron core and the U-shaped opening of the second U-shaped iron core can generate a double-sharp-corner magnetic field in the N-S-N-S direction below the welding wire. The magnetic generation device can generate a composite magnetic field formed by superposing a double-sharp-corner magnetic field and an axial magnetic field after being electrified, more energy can be used for melting welding wires, the welding wire melting efficiency is improved, a molten pool can be driven to rotate, the purpose of stirring the molten pool is achieved, grains are refined, and the performance of a formed part is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of arc fuse additive manufacturing, and in particular relates to a magnetron arc fuse additive manufacturing system and a manufacturing method thereof. Background Art

[0002] High-end equipment manufacturing industries such as advanced rail transit are strategic industries that provide technical equipment for my country's basic industries. However, in these industries, traditional casting, plastic forming, machining and other complex manufacturing processes still dominate, which has stimulated an urgent demand for flexible, low-cost, short-cycle and high-efficiency arc fuse additive manufacturing technology.

[0003] Arc fuse additive manufacturing refers to an additive manufacturing method that uses arc as a heat source to weld metal wire layer by layer to manufacture solid parts. CMT technology, or cold metal transfer technology, has the advantages of extremely low heat input, low flyback, and stable process. It is currently regarded as one of the most promising means of arc fuse additive manufacturing.

[0004] However, as the height of the deposit increases, the heat accumulation problem causes the metal pool to become less stable at higher layers and tends to flow toward the side walls, which in turn affects the surface forming accuracy and increases the cost of subsequent milling. In addition, during the arc fuse additive manufacturing process of thin-walled parts, if the arc width is greater than the width of the deposited part, part of the arc heat will act on the side walls of the deposited part, reducing the arc energy utilization rate and easily making the deposited metal grains coarse, affecting its performance. In addition, for aluminum alloy additive manufacturing, pores are one of the most common metallurgical defects, which seriously restricts the mechanical properties and quality of components. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to provide a magnetically controlled arc fuse additive manufacturing system and a manufacturing method thereof. On the one hand, a pulsed "sharp angle + axial" composite magnetic field is used to control the thermal distribution of the arc, realize arc energy redistribution and control the flow of the molten pool. On the other hand, based on molten pool image detection and deep learning, the geometric shape of the molten pool is predicted in real time and the magnetic field parameters are feedback-controlled, ultimately achieving the improvement of the geometric accuracy of the additive component and the improvement of the organizational performance.

[0006] The technical solution of the present invention is: a magnetically controlled arc fuse additive manufacturing system, comprising: The CMT welding machine has a welding gun with a welding wire at the lower end thereof; The magnetic generating device comprises: an electromagnetic generating part, which comprises: an annular iron core, which is a tubular structure, and is divided into an upper section, a middle section, and a lower section from top to bottom, and the annular iron core is sleeved on the outer side of the welding gun; a coil, which is wound on the annular iron core and is located in the middle section; a first U-shaped iron core, which is arranged on the annular iron core with a U-shaped mouth facing downward, and is located in the upper section; a second U-shaped iron core, which is arranged on the annular iron core with a U-shaped mouth facing downward, and is located in the lower section, and the opening of the second U-shaped iron core is flush with the opening of the first U-shaped iron core; a power supply part, which is electrically connected to the coil, and when the coil is energized, the annular iron core will generate an axial magnetic field along the axis direction of the welding wire at the welding wire, and the U-shaped mouth of the first U-shaped iron core and the U-shaped mouth of the second U-shaped iron core will generate a double-pointed magnetic field in the NSNS direction below the welding wire; The molten pool image detection device comprises: an image acquisition unit fixed on the welding gun, used for real-time acquisition of molten pool image information; a control unit connected to the image acquisition unit, used for storing and identifying molten pool image information and predicting the shape of the molten pool; the control unit is connected to the power supply unit, used for transmitting a signal to the power supply unit to realize feedback control of the magnetic field; The computer is connected to the CMT welding machine, the power supply unit, the image acquisition unit and the control unit respectively, and is used to realize the synchronous control of the image acquisition unit and the power supply unit during the arc additive process.

[0007] Furthermore, threads are provided on the upper outer wall and the lower outer wall of the annular core; The first U-shaped iron core and the second U-shaped iron core are both provided with mounting rings, which are threadedly mounted on the annular iron core through the mounting rings.

[0008] Furthermore, the angle between the plane where the U-shaped opening of the first U-shaped iron core is located and the plane where the U-shaped opening of the second U-shaped iron core is located is 30°-90°.

[0009] Furthermore, the number of turns of the coil is 100-400 turns.

[0010] Furthermore, the annular core, the first U-shaped core, and the second U-shaped core are all made of electrical pure iron.

[0011] Furthermore, the power supply unit is an excitation power supply, which is used to pass a pulse excitation current with a peak value of 1A-50A, a base current of 0A, and a pulse frequency of 1Hz-100Hz into the coil.

[0012] A method for additive manufacturing of a magnetron arc fuse, using the manufacturing system to additively manufacture the arc fuse, specifically comprising the following steps: S1. According to the shape of the arc fuse additive to be manufactured, the size of the cladding layer is set and appropriate additive process parameters are selected; S2. Set welding parameters on the CMT welding machine; S3, based on the computer, the additive process, the magnetic generation process and the image acquisition are synchronously started, and the arc fuse additive manufacturing starts to be started; S4. During the manufacturing process, the size of the cladding layer is monitored and feedback controlled in real time using a molten pool image detection device until the manufacturing is completed.

[0013] Furthermore, the step S4 is specifically as follows: S4-1, collecting molten pool images under different excitation parameters through a molten pool image detection device, and collecting the cladding layer width and excitation parameters corresponding to the molten pool images; S4-2. Using the control unit to construct a data set of multimodal information fusion based on the data collected in S4-1, and establish a deep learning-based cladding layer height regression prediction neural network model, train and optimize the cladding layer height regression prediction neural network model, and obtain a deep learning neural network model; S4-3. Input the real-time collected molten pool image and excitation parameter data into the deep learning neural network model obtained in S4-2 to obtain the cladding layer height. Through the proportional integral differential control algorithm, adjust the excitation parameters to achieve real-time, closed-loop cladding layer height control.

[0014] Compared with the prior art, the present invention has the following beneficial effects: After the magnetic generating device proposed in the present invention is powered on, it can generate a composite magnetic field formed by the superposition of a double-angle magnetic field and an axial magnetic field. Among them, the double-angle magnetic field can change the traditional circular arc into an elliptical shape, realize unidirectional compression of the arc, adapt to the width of thin-walled parts, avoid excessive arc energy heating the side wall, and increase the arc energy density, use more energy to melt the welding wire, and improve the melting efficiency of the welding wire. At the same time, the axial magnetic field can make the arc rotate in a circumferential direction, thereby driving the molten pool to rotate, achieve the purpose of stirring the molten pool, and then refine the grains and improve the performance of the formed part. By matching the pulse welding current and the pulse excitation current, that is, the excitation current is 0 at the base value welding current, there is no external magnetic field, and the arc interface is circular; the magnetic field is applied at the peak welding current stage, and the arc shrinks unidirectionally, so that it produces an inward arc shear force on the molten pool, promotes the inward flow of the molten pool, and avoids the molten pool flowing to the side wall. At the same time, based on the molten pool image detection, the height of the cladding layer is predicted in real time, and the excitation power supply is feedback-controlled to adjust the current size and frequency of the excitation power supply to achieve the control of the height of the cladding layer. In summary, this method can solve the problems of poor molten pool stability and coarse grains caused by high deposition height, thereby improving the efficiency, precision and performance of additive manufacturing of CMT arc fuses.

[0015] The present invention can be produced in conjunction with conventional CMT arc fuse additive manufacturing equipment, does not require modification of existing welding equipment or addition of a composite heat source, is low in cost, and is easy to modify. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the overall structure of the present invention; Figure 2 It is a schematic diagram of the local structure of the present invention; Figure 3 It is a structural schematic diagram of the magnetic generating device of the present invention; Figure 4 It is a structural schematic diagram of the annular core and the coil of the present invention; Figure 5 is a schematic structural diagram of a first U-shaped iron core of the present invention; Figure 6 is a schematic structural diagram of a second U-shaped iron core of the present invention; Figure 7 Schematic diagram of the direction of the magnetic field generated by the magnetic head of the magnetic generating device of the present invention Figure 8 Comparison of arc force diagrams in cross section with and without external composite magnetic field. Fig. 9 Comparison of arc force diagrams perpendicular to the cladding direction with and without an external composite magnetic field. Fig.10 This is a structural diagram of the deep learning neural network of the present invention; Fig.11 This is the structural diagram of the PID closed-loop control algorithm of the present invention.

[0017] Among them, 1-CMT welding machine, 10-welding gun, 100-welding wire, 2-magnetic generating device, 21-electromagnetic generating part, 211-annular iron core, 212-coil, 213-first U-shaped iron core, 214-second U-shaped iron core, 22-power supply part, 3-molten pool image detection device, 31-image acquisition part, 32-control part, 4-computer, 5-shielding gas part. DETAILED DESCRIPTION

[0018] Combine the following Figures 1 to 11 , the specific implementation methods of the present invention are described in detail. In the description of the present invention, it should be understood that the terms "center", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention.

[0019] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features; in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0020] It should be noted that the circuit connections involved in the present invention all adopt conventional circuit connection methods and do not involve any innovation.

[0021] Example like Figure 1 A magnetically controlled arc fuse additive manufacturing system is shown, comprising a CMT welding machine 1, a magnetic generating device 2, a molten pool image detection device 3, a computer 4 and a shielding gas unit 5.

[0022] The CMT welding machine 1 comprises a welding gun 10 , and a welding wire 100 is disposed at the lower end of the welding gun 10 .

[0023] The magnetic generating device 2 includes an electromagnetic generating unit 21 and a power supply unit 22. Figure 2 , Figure 3 As shown, the electromagnetic generating part 21 includes an annular core 211, a coil 212, a first U-shaped core 213 and a second U-shaped core 214. The annular core 211 is a tubular structure, which is divided into an upper section, a middle section and a lower section from top to bottom. The annular core 211 is sleeved on the outside of the welding gun 10. The coil 212 is wound on the annular core 211 and is located in the middle section. The U-shaped mouth of the first U-shaped core 213 is downwardly arranged on the annular core 211 and is located in the upper section. The U-shaped mouth of the second U-shaped core 214 is downwardly arranged on the annular core 211 and is located in the lower section. The opening of the second U-shaped core 214 is flush with the opening of the first U-shaped core 213. The power supply unit 22 is electrically connected to the coil 212. When the coil 212 is energized, the annular core 211 generates an axial magnetic field along the axis of the welding wire 100 at the welding wire 100. The magnetic field can rotate the arc. The U-shaped mouth of the first U-shaped core 213 and the U-shaped mouth of the second U-shaped core 214 generate a double-cusp magnetic field in the NSNS direction below the welding wire 100. The magnetic field acts on the arc and compresses the arc in one direction. The effective length of the first U-shaped core 213 is longer than that of the second U-shaped core 214.

[0024] like Figure 1As shown, the molten pool image detection device 3 includes an image acquisition unit 31 and a control unit 32. The image acquisition unit 31 is fixed on the welding gun 10 and is used to acquire the molten pool image information in real time. The control unit 32 is connected to the image acquisition unit 31 and is used to store and identify the molten pool image information and predict the molten pool shape; the control unit 32 is connected to the power supply unit 22 and is used to send a signal to the power supply unit 22 to realize feedback control of the magnetic field. In this embodiment, the image acquisition unit 31 adopts a commercially available high-speed camera.

[0025] like Figure 1 As shown, the computer 4 is respectively connected to the CMT welding machine 1, the power supply unit 22, the image acquisition unit 31, and the control unit 32, and is used to realize the synchronous control of the image acquisition unit 31 and the power supply unit 22 during the arc additive process; wherein, the computer 4 is a commercially available desktop computer.

[0026] The shielding gas section 5 is used to provide shielding gas to the molten pool. Preferably, the shielding gas section 5 includes a gas source and a gas delivery system. In this embodiment, the gas source is an argon gas cylinder, which provides argon gas to the molten pool through the gas delivery system, wherein the gas delivery system is a delivery pipeline, a valve and a nozzle, all of which are commercially available products. The nozzle is located near the welding gun 10 and is aligned with the welding gun 10. The delivery pipeline is used to connect the nozzle and the argon gas cylinder. The valve is arranged on the delivery pipeline for on-off control of the pipeline. This feature is a conventional technical means in the field and will not be described in detail here.

[0027] like Figure 5 , Figure 7 As shown, the magnetic poles of the four magnetic heads at the U-shaped mouth of the first U-shaped iron core 213 and the U-shaped mouth of the second U-shaped iron core 214 are NSNS, and this magnetic field is called a sharp-angle magnetic field. The magnetic pole of the magnetic head of the middle annular iron core 211 is N, and the direction is vertically downward along the axis of the annular iron core 211, which is called an axial magnetic field.

[0028] like Figure 7 , Figure 8 As shown, the interaction between the sharp-angle magnetic field and the current in the arc generates the Lorentz force, which can compress the arc in one direction, thereby adapting to the thickness of the additive part in the additive manufacturing process, avoiding excessive arc heat from acting on the side wall of the additive part, and also allowing more heat to melt the welding wire 100, thereby improving the arc energy utilization rate. At the same time, the Lorentz force generated by the interaction between the axial magnetic field and the radial current in the arc can cause the arc to rotate, thereby driving the molten pool to rotate, achieving the purpose of stirring the molten pool and refining the grains. In addition, as Fig. 9 As shown, the pulse magnetic field stimulated by the pulse excitation current compresses the arc periodically, thereby generating an arc shear force on the surface of the molten pool pointing to the inside of the arc, which can restrain the molten pool, promote the inward flow of the molten pool, and prevent the liquid metal in the molten pool from flowing to the side wall.

[0029] Preferably, Figure 4As shown, threads are provided on the upper outer wall and the lower outer wall of the annular core 211. The first U-shaped core 213 and the second U-shaped core 214 are both provided with mounting rings, which are mounted on the annular core 211 through the mounting rings.

[0030] Preferably, Figure 5 As shown, the angle between the plane where the U-shaped opening of the first U-shaped core 213 is located and the plane where the U-shaped opening of the second U-shaped core 214 is located is 30°-90°. In this embodiment, it is 90°.

[0031] Preferably, the number of turns of the coil 212 is 100-400 turns, and in this embodiment, it is 200 turns.

[0032] Preferably, the annular core 211, the first U-shaped core 213, and the second U-shaped core 214 are all made of electrical pure iron, which has the advantages of high magnetic permeability and low magnetic loss.

[0033] Preferably, the power supply unit 22 is an excitation power supply, which is used to pass a pulse excitation current with a peak value of 1A-50A, a base current of 0A, and a pulse frequency of 1Hz-100Hz into the coil 212.

[0034] A method for additive manufacturing of a magnetron arc fuse, using the manufacturing system proposed in this embodiment to perform additive manufacturing of an arc fuse, specifically comprises the following steps: S1. According to the shape of the arc fuse additive to be manufactured, the size of the cladding layer is set and appropriate additive process parameters are selected; S2. Set welding parameters on CMT welding machine 1.

[0035] S3, turn on the power supply unit 22, set the excitation parameters, the excitation current peak value is 1A-50A, the base current is 0A, and the pulse frequency is 1Hz-100Hz. Turn on the high-speed camera and adjust the relevant parameters of the high-speed camera. Turn on the computer 4 to achieve the synchronous start of the additive process, the magnetic generation process and the image acquisition, and the arc fuse additive manufacturing starts and manufactures.

[0036] S4. During the manufacturing process, the molten pool image detection device 3 is used to monitor and feedback-control the size of the cladding layer in real time until the manufacturing is completed.

[0037] Preferably, Fig.11 As shown, step S4 is specifically as follows: S4-1. Collect the molten pool images under different excitation parameters through the molten pool image detection device 3, and collect the cladding layer width and excitation parameters corresponding to the molten pool images.

[0038] S4-2, using the control unit 32 to construct a data set of multimodal information fusion based on the data collected in S4-1, and establish a neural network model for regression prediction of cladding layer height based on deep learning, train and optimize the neural network model for regression prediction of cladding layer height, and obtain the following: Fig.10 The deep learning neural network model shown.

[0039] S4-3. Input the real-time collected molten pool image and excitation parameter data into the deep learning neural network model obtained in S4-2 to obtain the cladding layer height. Through the proportional integral differential control algorithm, adjust the excitation parameters to achieve real-time, closed-loop cladding layer height control.

[0040] Preferably, the excitation parameters in step S4-1 include an excitation current value and an excitation current frequency.

[0041] Preferably, in step S4-2, the data set includes a molten pool image, an excitation current value, an excitation current frequency, and a cladding layer height, wherein the molten pool image is used to extract features, the excitation current value and frequency provide information related to arc control, and the cladding layer height is used as a target label. The deep learning neural network model is based on an improved ResNet-34 residual network architecture, which includes a ResNet-34 residual network, a fully connected layer branch, and a multimodal information feature fusion layer; the model uses the molten pool image and excitation parameters in the data set as input, wherein the ResNet-34 residual network extracts high-dimensional features from the molten pool image, and the fully connected layer branch processes the normalized excitation parameters to generate a representation that matches the image features. The two parts of the feature are deeply fused through a multimodal feature fusion layer, and the fully connected layer and the regression layer are sequentially connected after the feature fusion layer to output the height of the cladding layer.

[0042] Preferably, the data set for multimodal information fusion in S4-2 is produced based on the cladding layer height-annotated molten pool image and excitation parameter data.

[0043] Preferably, in S4-2, the mean square error MSE is used to predict the actual height and width of the cladding layer, and the calculation formula is: , where is the predicted value of the cladding layer height regressed by the model, yi is the actual value, n is the number of samples, and i is the data sample of each one.

[0044] Preferably, the excitation parameters in S4-3 are controlled by the excitation power supply, and the excitation power supply is connected to the computer through a control cabinet.

[0045] S4-1 specifically includes the following steps: S4-11: Adjust the excitation current and frequency.

[0046] S4-12: With the welding wire 100 as the center, adjust the ROI size of the image acquisition unit 31 to 512*256 pixels; set the working exposure of the image acquisition unit 31 to 200us and the working frequency to 1000Hz.

[0047] S4-13: After the welding of the current layer is completed and cooled, the current cladding layer is scanned with a 3D scanner to obtain the relationship between the layer height and the position of the cladding layer. The height difference between the positions of two adjacent layers of the same cladding layer is the height increment of the current cladding layer position.

[0048] S4-14: The position of the cladding layer is obtained through the molten pool image time, and the corresponding relationship between the molten pool image, the excitation parameters and the cladding layer height is further obtained.

[0049] S4-2 specifically includes the following steps: S4-21: The molten pool image is scaled to 256*128 pixels, annotated according to the correspondence between the molten pool image, excitation parameters and cladding layer height, and a multimodal information fusion data set is produced.

[0050] S4-22: Use the ResNet-34 residual network structure to regress the cladding layer height, remove the classification layer in the ResNet-34 residual network structure, and add a fully connected layer.

[0051] S4-3 specifically includes the following steps: S4-31: The optimized model in step S4-2 is embedded in the real-time system, and the current layer height is predicted using the real-time collected molten pool image and excitation parameters.

[0052] S4-32: Select the excitation current and excitation frequency as the input of the magnetron arc additive manufacturing monitoring system, select the cladding layer height as the output of the magnetron arc additive manufacturing monitoring system, and fit the layer height group with the best effect through step response experiment: The first-order transfer function of the excitation current and excitation frequency with the layer height is: , where represents the excitation current gain, Indicates the excitation frequency gain; and represents the time constant of the first-order system; s represents the Laplace transform differential operator.

[0053] S4-33: According to the target layer height and the actual floor height predicted by the neural network , calculate the real-time error : .

[0054] S4-34: Using a First-Order Transfer Function and error signal , design PID controller to adjust the excitation current and excitation frequency: The PID control output formula is: , where u(t) represents the control output signal, which is transmitted to the excitation power supply control cabinet; represents the proportional gain of PID, Represents the integral gain of PID, Indicates the differential gain of PID; Represents the time constant of the controlled object.

[0055] S4-35: Use the PID controller to calculate the adjustment value u(t) of the excitation current and excitation frequency, which is transmitted to the excitation power supply control cabinet by the computer, and the cladding layer height is fed back in real time: ,in, , where and Represent the basic excitation current and basic excitation frequency respectively; and Indicates the dynamic adjustment value calculated by PID control.

[0056] The specific models of the above electronic components are not specially specified, and ordinary products available on the market can be selected as long as they can meet the use requirements of the present invention.

[0057] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and does not limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims

1. A magnetically controlled arc fuse additive manufacturing system, characterized in that: include: The CMT welding machine (1) comprises a welding gun (10), wherein a welding wire (100) is provided at the lower end of the welding gun (10); The magnetic generating device (2) comprises: an electromagnetic generating part (21), which comprises: an annular iron core (211), which is a tubular structure and is divided into an upper section, a middle section, and a lower section from top to bottom, the annular iron core (211) being sleeved on the outside of a welding gun (10); a coil (212), which is wound around the annular iron core (211) and is located in the middle section; a first U-shaped iron core (213), which is arranged on the annular iron core (211) with its U-shaped opening facing downwards and is located in the upper section; and a second U-shaped iron core (214), which is arranged on the annular iron core (211) with its U-shaped opening facing downwards. The upper portion is located in the lower section, and the opening of the second U-shaped iron core (214) is flush with the opening of the first U-shaped iron core (213); the power supply unit (22) is electrically connected to the coil (212), and when the coil (212) is energized, the annular iron core (211) generates an axial magnetic field along the axis direction of the welding wire (100) at the welding wire (100), and the U-shaped mouth of the first U-shaped iron core (213) and the U-shaped mouth of the second U-shaped iron core (214) generate a double-pointed magnetic field in the NSNS direction below the welding wire (100); A molten pool image detection device (3) comprises: an image acquisition unit (31) for acquiring molten pool image information in real time; a control unit (32) connected to the image acquisition unit (31) and used to store and identify molten pool image information and predict the shape of the molten pool; the control unit (32) is connected to the power supply unit (22) and used to transmit a signal to the power supply unit (22) to achieve feedback control of the magnetic field; The computer (4) is respectively connected to the CMT welding machine (1), the power supply unit (22), the image acquisition unit (31), and the control unit (32), and is used to realize synchronous control of the image acquisition unit (31) and the power supply unit (22) during the arc material addition process.

2. A magnetically controlled arc fuse additive manufacturing system as claimed in claim 1, characterized in that: The upper outer wall and the lower outer wall of the annular iron core (211) are both provided with threads; The first U-shaped iron core (213) and the second U-shaped iron core (214) are both provided with a mounting ring, and are threadedly mounted on the annular iron core (211) via the mounting ring.

3. A magnetically controlled arc fuse additive manufacturing system as claimed in claim 2, characterized in that: The angle between the plane where the U-shaped opening of the first U-shaped iron core (213) is located and the plane where the U-shaped opening of the second U-shaped iron core (214) is located is 30°-90°.

4. A magnetically controlled arc fuse additive manufacturing system as claimed in claim 1, characterized in that: The number of turns of the coil (212) is 100-400 turns.

5. The magnetically controlled arc fuse additive manufacturing system according to claim 1, characterized in that: The annular iron core (211), the first U-shaped iron core (213), and the second U-shaped iron core (214) are all made of electrical pure iron.

6. The magnetically controlled arc fuse additive manufacturing system according to claim 1, characterized in that: The power supply unit (22) is an excitation power supply, which is used to supply a pulse excitation current with a peak value of 1A-50A, a base current of 0A, and a pulse frequency of 1Hz-100Hz to the coil (212).

7. A method for additive manufacturing of a magnetron arc fuse, characterized in that: Using the manufacturing system described in any one of claims 1 to 6 to perform arc fuse additive manufacturing specifically comprises the following steps: S1. According to the shape of the arc fuse additive to be manufactured, the size of the cladding layer is set and appropriate additive process parameters are selected; S2. Setting welding parameters on the CMT welding machine (1); S3, based on the computer (4), the additive process, the magnetic generation process and the image acquisition are started synchronously, and the arc fuse additive manufacturing starts and the manufacturing is carried out; S4. During the manufacturing process, the size of the cladding layer is monitored and feedback-controlled in real time using the molten pool image detection device (3) until the manufacturing is completed.

8. A method for additive manufacturing of a magnetron arc fuse according to claim 7, characterized in that: The step S4 is specifically as follows: S4-1, collecting molten pool images under different excitation parameters by means of a molten pool image detection device (3), and collecting the cladding layer width and excitation parameters corresponding to the molten pool images; S4-2, using the control unit (32) to construct a data set of multimodal information fusion based on the data collected in S4-1, and establish a deep learning-based cladding layer height regression prediction neural network model, and train and optimize the cladding layer height regression prediction neural network model to obtain a deep learning neural network model; S4-3. Input the real-time collected molten pool image and excitation parameter data into the deep learning neural network model obtained in S4-2 to obtain the cladding layer height. Through the proportional integral differential control algorithm, adjust the excitation parameters to achieve real-time, closed-loop cladding layer height control.

Images